Using longitudinal data from the 1997 National Longitudinal Study of Youth (NLSY-97), the research outlined in this proposal will extend previous research on the relationship between military service and troubled behaviors (antisocial, violent, criminal, or dangerous behaviors). The work being proposed will extend beyond the existing literature by explicitly considering how voluntary military service surrounding both sides of 9/11 is linked to numerous measures of troubled behavior. We broaden the concept of a troubled background to include numerous indicators of self-reported behavioral problems (e.g., arrests, sentencing, gang participation, violent behavior) and substance use (e.g., use of alcohol and illegal drugs) on enlistment. While Sampson and Laub (1996) present evidence in support of a rehabilitative effect for veterans of World War II, other research has been less consistent. For example, some research has found that veterans of later eras are more likely than non-veterans to use illegal drugs and engage in criminal activities (Bouffard, 2006; Galiani, Rossi, and Schargrodsky, 2006; Wright, Carter, and Cullen, 2005), while Bouffard (2003) reports either no relationship or a decrease in criminal behavior among veterans. In part, the variation in results may stem from the different samples used and methodological weaknesses. The methodological weaknesses include failure to account for selectivity into the military (especially with respect to prior troubled behavior) and limited, inconsistent measures of troubled behavior. We correct these weaknesses by utilizing nationally representative data with repeated indicators of different troubled behaviors that allow estimation of fixed-effects models, reducing the influence of selectivity. We also determine whether the link between military service and subsequent troubled behavior varies according to pre-service characteristics and historical period (Bouffard, 2006). In our analyses, we pay particular attention to pre-service disadvantage as indicated by race and socioeconomic status and the nature of service as indicated by service occurring before or after 9/11. Finally, we consider how the links between military service and troubled behavior vary according to gender. PUBLIC HEALTH RELEVANCE: Using data taken from a nationally-representative data base, the proposed research will extend our knowledge of the relationship between military service and troubled behavior. The work being proposed will expand beyond existing research by explicitly considering how multiple dimensions of troubled behavior are related to military service. We will make use of longitudinal data and fixed-effects models to account for selectivity of recruits into military service.